Machine Learning Techniques for Solving Quantum Many-Body Problems (QuanticLearn)

Objectives [EN]

The main objectives of the project concern an efficient description of quantum many-body problems (QMB) using machine learning ML techniques, in particular using ANNs, and the design of quantum neuron devices. The QMB systems shall be approached from at least two perspectives. The first one is concerned with the description of many-body scattering wavefunctions, within the framework of scattering theory. In the second perspective, the transport problem is approached using DFT-NEGF. These two perspectives offer both an opportunity to investigate fundamental aspects regarding entanglement, but also a pragmatic way to treat technological relevant aspects. The concrete objectives of the project are:

1. Description of quantum many-body scattering processes, in particular 2-particle scattering, using ML techniques;
2. Construct an efficient framework based on DFT-NEGF approach to quantum transport using ML techniques (DFT-NEGF-ML) with applications to nanoelectronic devices;
3. Design of quantum neuron devices and learning schemes for QNNs.

Obiective [RO]

Principalele obiective ale proiectului privesc o descriere eficienta a problemelor multi-particula (QMB) folosind tehnici de invatare automata, in particular retele neurale artificiale, precum si design-ul de dispozitive neurale. Problemele multi-particula vor fi abordate din cel putin doua perspective. Prima perspectiva priveste descrierea functiilor de unda multi-particula in cadrul teoriei de imprastiere. In a doua perspectiva, problema de transport este abordata folosind tehnica DFT-NEGF. Aceste doua perspective ofera oportunitatea de a investiga probleme fundamentale pentru entanglement, oferind totodata o cale pragmatica de a aborda probleme cu relevanta tehnologica. Obiectivele concrete ale proiectului sunt:

1. Descrierea proceselor de imprastiere de tip multi-particula, in particular imprastieri bi-particula, folosind tehnici de invatare automata;
2. Construirea unui cadru eficient avand la baza tehnica de tip DFT-NEGF pentru transport cuantic, folosind tehnici de invatare automata cu aplicati ila dispozitive nanoelectronice;
3. Design-ul dispozitivelor de tip neuron artificial cuantic si schemelor de invatare pentru retele neurale cuantice (QNNs).